Available online at www.sciencedirect.com ScienceDirect Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 2 nd World Conference on Business, Economics and Management WCBEM 2013 Explaining the cross section of stock returns: A comparative study of the United States and Turkey Yigit Bora Senyigit *, Yusuf Ag Zirve University,Gaziantep 27260, Turkey Zirve University,Gaziantep 27260, Turkey Abstract Stock returns are affected by various accounting information. Based on literature review, we identify three independent variables: P/E ratio (price to earnings ratio), P/B ratio (price to book ratio), and D/E ratio (debt to equity ratio). Focusing on these variables, this research aims to provide significant evidence to assess the impacts of these independent variables on stock returns. We have collected the necessary data of companies listed in the US and Turkey markets, and ended up with 452 observations from Turkey and 588 observations from the United States. We calculate annual stock returns based on the average of the past 12 monthly returns for each company. We use the panel data method to effectively assess the relationship between variables. This study suggests that the explanatory power of independent variables is relatively high and statistically significant in explaning the cross-section of stock returns in the United States; however, it is not in Turkey. 2014 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer review under responsibility of Organizing Committee of BEM 2013. Keywords: Cross section of stock returns, market efficiency, panel data; 1. Introduction Relations between stock returns and accounting information in the United States have been studied extensively. We hope to contribute this debate making comparison between the United States as a developed equity market and Turkey as an emerging equity market. In this study, we apply OLS-Pooled and GLS-Random Effects models to estimate stock returns for the cross section data. We collected the necessary data about selected models of the companies. The annual stock return is calculated based on the average of the past 12 monthly returns for each company. Based on this explanations, the objective of this study is to present statistically significant evidence to monitor the effects of various variables such as P/E ratio (price to earnings ratio), P/B ratio (price to book ratio), and D/E ratio (debt to equity ratio) on stock returns through the regression analysis which is often used for prediction and equation development for predictive purposes. The remainder of this paper is organized as follows. Section 2 presents the literature review. Section 3 explains data and methodology. Section 4 discusses the findings. Finally, Section 5 concludes the study. * Corresponding Author: Yigit Bora Senyigit. Tel.: +90-342-211-6666 E-mail address: bora.senyigit@zirve.edu.tr 1877-0428 2014 The Authors. Published by Elsevier Ltd. Open access under CC BY-NC-ND license. Selection and peer review under responsibility of Organizing Committee of BEM 2013. doi:10.1016/j.sbspro.2013.12.466
328 Yigit Bora Senyigit and Yusuf Ag / Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 2. Literature Review Discussions of the theory of stock price behavior start with Markowitz (1952, 1959). The Markowitz (1959) discusses that investors only use two measures of portfolio s risk, return and standard deviation, to choose investments. After Markowitz s study, Sharpe (1964) and Lintner (1965) extend the work and develop the most recognized and widely used asset pricing model called the Capital Asset Pricing Model (CAPM). This model has similar assumptions as Markowitz s model in structuring portfolios. The CAPM also assumes that there is a linear risk return relationship. Recent research in empirical accounting and finance has shown that variables such as P/E ratio (price to earnings ratio), P/B ratio (price to book ratio), and D/E ratio (debt to equity ratio) have significant explanatory power for the variation in cross section of expected returns. In this section, early empirical works were discussed to understand the relationship between stock returns and these independent variables theoretically. Basically, the price to earnings ratio is defined as the price an investor is paying for $1 of the company's earnings. Basu (1977) found that stocks with low price-earnings ratios (P/Es) have higher average returns than stocks with high P/Es. Additionally, earnings-price ratios (the reciprocal of our P/E ratio) help out clarifies the cross-section of average returns on the U.S. stocks in tests that also contain size and market beta (Basu, 1983). The P/B or price to-book ratio is a basic measure of the relative value that the market places on a share of stock. Fama and French (1992) indicated that the book to market ratio (the reciprocal of our P/B ratio) has the strongest relation with expected stock returns in the United States. High book-to-market stocks have higher returns than low book-to-market stocks (Fama and French, 1995). Therefore, low P/B stocks generate higher returns than the market. Debt to equity ratio is defined as a measure of a company's financial leverage calculated by dividing its total liabilities by stockholders' equity. Bhandari (1988) reported that stock returns are positively related to debt-equity ratios (D/E). Over the past decade or so, the number of studies explaining the cross-section of stock returns in developed and emerging markets has proliferated. The literature shows that many firm characteristics are correlated with subsequent stock returns (Fama and French, 1992). Similar results are reported for several developed equity markets (Ferson and Harvey, 1997; Fama and French, 1998; Penman, et. al., 2006; Lewellen, 2013), as well as emerging equity markets (Bekaeert, et. al., 1997; Patel, 1998; Rouwenhorst, 1999; Aydogan and Gursoy, 2000; Akdeniz, et. al., 2000). Bekaert et. al.(1997) found that the behavior of emerging market returns contradicts significantly from the behavior of developed market return. The study indicated that while there are some similarities between the cross-sectional characteristics of emerging and developed equity market stock returns, emerging market strategies have to consider the special characteristics of these markets. Akdeniz et. al. (2000) examined significance of explanatory factors, namely book-to-market, size and earnings-to-price, on stock returns during the period of 1992-1996. They found that these variables have some explanatory power during the first period that covers between 1992 and 1995, however, no explanatory power during second period that covers between 1995 and 1998. 3. Data and Methodology The data of study consists of the companies listed in ISE-100 and the companies listed in S&P-100 for the period 2005-2011. We have collected data from Compustat database. After cleaning of data (eliminating outliers and missing values) and removing the financial companies from data set, the remaining 76 of ISE companies and 84 of S&P companies are included in the analysis. Table 1 and Table 2 present descriptive statistics of these data sets.
Yigit Bora Senyigit and Yusuf Ag / Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 329 Table 1. Descriptive Statistics Panel: Turkey Variable Observation Mean Std. Dev. Min Max SR 452 0.28 0.66-0.86 2.14 P/E 452 11.22 45.29-405.83 426.67 P/B 452 1.49 0.91 0 4.59 D/E 452 1.17 1.08 0.01 5.89 Table 2. Descriptive statistics Panel: the United States Variable Observation Mean Std. Dev. Min Max SR 588 0.08 0.35-0.80 3.37 P/E 588 18.66 23.06-222.93 266.19 P/B 588 4.10 6.44-94.38 79.74 D/E 588 1.32 11.51-245.66 48.37 As indicated earlier, the purpose of the study is to investigate the indicators of stock returns and to find out how well these indicators can explain the variation in stock returns and whether explanatory power of these indicators varies by country. In this respect, the model of our study is set as follows: SR i = β 0 + β 1 P/E i + β 2 P/B i + β 3 D/E i + + ε i where, SRi is annual stock returns, P/Ei is price to earnings ratio, P/Bi is price to book ratio, D/Ei is debt to equity ratio and is a set of dummy variables that identify the variances in stock returns by years. Given the fact that our data is a cross-sectional in nature, we use the following regression techniques: OLS with clustered standard errors and GLS with random effects. Application of two different techniques would help us to check the robustness of parameter estimates. Regarding the possible multicollinearity issues, we check the correlation matrix of variables. Panel-Turkey Table 3. Correlation Matrix Panel-the United States SR P/E P/B D/E SR P/E P/B D/E SR 1 1 P/E 0.0248 1 0.0775 1 P/B 0.0029 0.2106 1 0.0315 0.1521 1 D/E -0.0343-0.0091 0.0733 1-0.0956 0.0226 0.6976 1 As seen in Table 3, the correlation matrix does not include any pair of two independent variables that has a correlation more than 90%. Hence, there is no multicollinearity between variables. 4. Discussion of Findings In this section, the findings of the empirical results and statistical analysis are presented. Table 4 lists parameter estimates from two different models.
330 Yigit Bora Senyigit and Yusuf Ag / Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 Table 4. Panel: Turkey Variables OLS-Pooled GLS-Random effects P/E 0.00000543 0.00000543 [0.00000849] [0.00000849] P/B -0.0000303-0.0000303 [0.0001275] [0.0001275] D/E -0.005795-0.005795 [0.0042384] [0.0042384] Year:2005 0.9980991*** 0.9980991*** [0.0866574] [0.0866574] Year:2006 0.4658091*** 0.4658091*** [0.0580896] [0.0580896] Year:2007 0.7013759*** 0.7013759*** [0.0683779] [0.0683779] Year:2009 1.730443*** 1.730443*** [0.1181223] [0.1181223] Year:2010 0.9317156*** 0.9317156*** [0.0823542] [0.0823542] Year:2011 0.3227651*** 0.3227651*** [0.0645772] [0.0645772] R-square 0.5042 0.5042 Robust standard errors are in brackets. *** p<0.01 The estimated parameters of OLS-Pooled and GLS-Random Effects models are same. The both regressions feature R-square of 0.50, meaning that independent variables in the model explain 50% of the variations in stock returns and there are other factors affecting the stock returns. Price to earnings ratio has positive relationship and price to book ratio and debt to equity ratio have negative relationship with stock returns. However, none of the independent variables are significant. Regarding the annual variations in stock returns, it seems that stock returns are increased and then decreased after global crisis in 2008. It appears that the economic shock in 2008 had a short-term positive effect in stock returns in 2009 but in following years, average stock return is dropped to the lower level than that in pre-2008. This shows that Turkish stock market is still under global crisis effect. Table 5 lists the estimated parameters of Panel-the United States from two different models.
Yigit Bora Senyigit and Yusuf Ag / Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 331 Table 5. Panel: the United States Variables OLS-Pooled GLS-Random effects P/E 0.0006009 0.0006009 [0.0006313] [0.0006313] P/B 0.0094858*** 0.0094858*** [0.0032251] [0.0032251] D/E -0.006659*** -0.006659*** [0.0011498] [0.0011498] Year:2005 0.3428123*** 0.3428123*** [0.0405947] [0.0405947] Year:2006 0.3331744*** 0.3331744*** [0.0324446] [0.0324446] Year:2007 0.4476662*** 0.4476662*** [0.0441211] [0.0441211] Year:2009 0.5915979*** 0.5915979*** [0.076556] [0.076556] Year:2010 0.4269221*** 0.4269221*** [0.0388317] [0.0388317] Year:2011 0.3633784*** 0.0273706*** [0.0273706] [0.0263074] R-square 0.2559 0.2559 Robust standard errors are in brackets. *** p<0.01 As seen in Table 5, the estimated parameters of OLS-pooled and GLS-random effects are same. Price to book ratio have positive and significant (at 1%) relationship with stock returns and debt to equity ratio has negative and significant (at 1%) relationship with stock returns. However, price to earnings has positive but insignificant relationship with stock returns. Both regressions feature R-square of 0.26 which is lower than that of Panel- Turkey. This means that there are a significant number of other factors that affect the variations in stock returns. Comparing Turkey and US panels, the explanatory power of variables are higher in Turkey panel but it has no ground due to insignificant relationships of variables. When looking at the annual variations of stock returns, it appears that the average stock return of post-2008 is higher than that of pre-2008. This shows that the United States stock market is slowly recovering from the global crisis effect. 5. Conclusion Aiming to present statistically significant evidence to monitor the effects of various variables such as; price to earnings ratio, price to book ratio, and debt to equity ratio through the regression analysis, we concluded that all three independent variables have an impact on dependent variable, which is stock returns, in the United States. However, we could not find statistically significant relationship between dependent variable and independent
332 Yigit Bora Senyigit and Yusuf Ag / Procedia - Social and Behavioral Sciences 109 ( 2014 ) 327 332 variables in Turkey data set. Although our research does not offer any clue for the underlying reasons, we nevertheless suggest some insights. Emerging markets are different from developed markets in terms of their nature and inherent characteristics. Emerging markets are more volatile than develop markets. Therefore, it is understandable that the explanatory power of independent variables is relatively high and statistically significant in explaning the cross-section of stock returns in the United States; however, it is not in Turkey. References Akdeniz, L., Salih, A., and Aydogan, K. (2000). Cross section of expected returns in ISE. Russian and East European Finance and Trade, 36(5), 6-26. Aydogan, K., and Gursoy, G. (2000). P/E and price-to-book ratios as predictors of stock returns in emerging equity markets. Emerging Markets Quarterly, 4(4), 60-67. Basu, S. (1977). The investment performance of common stocks in relation to their price-earnings ratios: A test of the efficient market hypothesis. Journal of Finance, 32, 663-82. Basu, S. (1983). The relationship between earnings yield, market value, and return for NYSE common stocks: Further evidence. Journal of Financial Economics, 12, 129-156. Bekaert, G., Erb, C. B., Harvey, C. R. and Viskanta, T. E. (1997).The Cross-Sectional Determinants of Emerging Market Equity Returns. (pp.221-272), in Carman, Peter, Ed., Quantitative Investing for the Global Markets, Glenlake. Bhandari, L.C. (1988.) Debt/Equity ratio and expected common stock returns: Empirical evidence. The Journal of Finance, 43, 507-28. Fama, E.F. and French, K.R. (1992). The cross-section of expected stock returns. The Journal of Finance, 47, 427-66. Fama, E.F. and French, K.R. (1996). Multifactor explanations of asset pricing anomalies. The Journal of Finance, 51, 55-84. Lewellen, J. (2013). The cross section of expected stock returns. Working paper, Dartmouth College. Lintner, J. (1965). The valuation of risk assets and the selection of risky investments in stock portfolios and capital budgets, Review of Economics and Statistics, 47, 13-37. Markowitz, H. (1952). Portfolio selection. The Journal of Finance, 7, (1), 77-91. Markowitz, H. (1959). Portfolio selection, efficient diversification of investments. New York: John Wiley and Sons, Inc. Penman, S. H., Richardson, S. A. and Tuna, I. (2007), The book-to-price effect in stock returns: Accounting for leverage. Journal of Accounting Research, 45, 427 467. Sharpe, W.F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk. The Journal of Finance, 19, 425-442.